基于VMD-BP-GA模型的脆弱航段船舶短时交通流预测

Short-term Ship Traffic Flow Prediction in Vulnerable Segments Based on VMD-BP-GAModel

【目的】针对繁忙航段船舶交通流易受外界环境扰动的难题,提出一种可用于识别船舶交通流脆弱性的预测模型,旨在通过脆弱性辨识,确定最薄弱的航段。【方法】首先采用变分模态分解(VMD)模型将船舶交通流参数序列分解为多个模态分量,然后结合反向传播神经网络(BP)和遗传算法(GA),通过构建约束模型并不断更新各个分量的中心和带宽,实现单个分量的预测,通过应用VMD-BP-GA模型对船舶交通流进行精准预测,并验证其合理性和有效性。【结果】在繁忙航段,本研究提出的VMD-BP-GA模型精准预测船舶交通流脆弱性的方法,相较于传统模型表现出更低的预测误差值,其中在航段流量预测方面,本研究模型的平均绝对误差(MAE)最低达到2.095%,均方根误差(RMSE)最低达到2.610%,平均百分比误差(MAPE)最低达到2.114%;在航段密度预测方面,本研究模型的MAE、RSME、MAPE最低分别为0.129%、0.162%、2.112%;并实现了时空两个维度的船舶交通流预测。【结论】本研究模型成功实现对船舶交通流脆弱性的识别和最薄弱航段的确定,具有高效的预测性能,能够精准并快速地预测船舶交通流,可为船舶通航安全保障提供了理论和实践指导。

【Objective】Against the challenge of ship traffic flows being susceptible to external environmental disturbances in busy searoutes,a predictive model for identifying the vulnerability of ship traffic flows is proposed,aiming to determine the weakest segments through vulnerability identification.【Method】Firstly,the Variational Mode Decomposition(VMD)model was employed to decompose the parameters of ship traffic flows into multiple modal components.Then,combining Back-Propagation Network(BP)and Genetic Algorithm(GA),a constrained model was constructed to continuously update the center and bandwidth of each component,achieving the prediction of individual components.Through the application of VMD-BP-GA model,ship traffic flows were accurately predicted,and the rationality and effectiveness of the model were validated.【Result】Based on the VMD-BP-GA model,this study proposes a method for accurately predicting the vulnerability of ship traffic flow.In busy shipping routes,this method performs better in terms of error metrics compared to traditional models,with the lowest mean absolute error(MAE)reaching 2.095%,root mean square error(RMSE)reaching 2.610%,and mean percentage error(MAPE)reaching 2.114%for the ship traffic flow prediction model in the routes.In terms of segment density prediction,MAE,RMSE and MAPE of this method reach the lowest of 0.129%,0.162%and 2.112%respectively.Moreover,it achieves predictions in both spatial and temporal dimensions.【Conclusion】The model successfully identifies the vulnerability of ship traffic flow and determines the most vulnerable shipping route.It demonstrates efficient predictive performance,enabling accurate and rapid prediction of ship traffic flow,thus providing theoretical and practical guidance for ensuring maritime safety.